PhD Student Researcher - xLab for Safe Autonomous SystemsTeacher's Assistant: ESE6150 F1/10 Autonomous Racing Cars

Worked on developing a model-learning pipeline from raw ROS-bag data.

🞄 Implemented a ROS2 wrapper for a kinematic model-predictive controller (MPC) for a 1/10th scale autonomous car.🞄 Achieved a 400x speedup in control frequency using a sparse-matrix formulation and CVXPYgen.🞄 Calibrated a Qualisys Motion Capture system and collected datasets for controller validation.🞄 Implemented a novel weight-adaptation algorithm for Gaussian Process ensembling.

Designed and developed an April-tag vision system for the Franka Panda robot arms to detect and pick up blocks in the environment. Augmented the final project simulation with an emulated camera to simulate the camera detections of the april tags.Primary Duties:Holding office hours for students.Grade homework submissions.Setting up and directing the final Pick-and-Place competition.

• Participated in bi-weekly sprints on Jira, using Bitbucket and GiT for version control and code reviews.• Developed firmware for ARM and Tensilica Xtensa LX6 chips for audio and control applications.• FreeRTOS firmware development in an IoT development framework and an audio development framework.• Utilized SPI, I2C, I2S, and UART protocols to communicate with different peripherals.• Implemented the driver for the database system in the Roadie 3 guitar tuner using C and nrf5 SDK• Utilized test-driven development to rewrite our Bluetooth communication protocol, using C/C++ and gTest• Circuit design using KiCad.• Collaborated with the software development team to ensure proper connection to the mobile app and AWS.

Worked on Human-Robot interaction through Smart Assistants like Amazon's Alexa:• Developed a bidirectional cloud communication architecture that connects robots to Amazon’s Alexa using AWS IoT.• Used Amazon Lex and gTTs to enable robot vocal communication.• Designed and built a testing scenario for human subject testing using ROS and Gazebo.• Developed a desktop application as a comparative robot interaction method, using python and Tkinter.• Developed custom Gazebo models using Blender and XML for use in the testing scenario.• Wrote a draft for a research paper whose publishing is pending the collection of testing data from human‑subject testing.

Programming Languages Used: MATLAB, C, PythonTools Used: Keil μVision, Eclipse (C programming for Embedded Systems), Putty SSHProtocols Used: UART, SPI, UDP, TCP, MQTT, OTA (WiFi)• Redesigned the screen UI based on designer specifications, using C and the u8g2 library.• Implemented a discrete feed‑forward motor position controller to cut costs down, implemented in C for an ARM microprocessor.• Designed a custom UART communication protocol for real‑time debugging using C, with a MATLAB GUI for visualization.Skills Learned: Research & Development, Software Design, Quality Control (Software/Hardware), Digital Controller (Control Theory) Design and Implementation, Digital Communication, Screen (OLED) Interface programming, GUI Programming (MATLAB)

Worked on developing a novel sensor for detecting 100μm - 1000μm oil-slick thickness:. Conducted exhaustive literature review and market survey of available technologies.. Suggested and tested the feasibility of a imaging-based approach for detecting the oil thickness using lateral imaging: . System showed promise and plausibility. . The system was then tested for repeatability and accuracy under a controlled environment (lab).End Result:Retired from research as the project was handed to another group as their final year project. Currently on-hold as a reference for the team on any of the work we already did (papers, techniques, reuslts, etc...).

Teacher's Assistant for Three Courses;Introduction to Programming (Jan 2018 - April 2018)Electronics (August 2018 - December 2018)Electronic Circuits (January 2019 - April 2019)